Download Working Paper 177 - Grodecka & Kuralbayeva (opens in new window)

The price vs quantity debate: climate policy
and the role of business cycles
Anna Grodecka and Karlygash Kuralbayeva
January 2015
Centre for Climate Change Economics and Policy
Working Paper No. 201
Grantham Research Institute on Climate Change and
the Environment
Working Paper No. 177
The Centre for Climate Change Economics and Policy (CCCEP) was established
by the University of Leeds and the London School of Economics and Political
Science in 2008 to advance public and private action on climate change through
innovative, rigorous research. The Centre is funded by the UK Economic and Social
Research Council. Its second phase started in 2013 and there are five integrated
research themes:
1. Understanding green growth and climate-compatible development
2. Advancing climate finance and investment
3. Evaluating the performance of climate policies
4. Managing climate risks and uncertainties and strengthening climate services
5. Enabling rapid transitions in mitigation and adaptation
More information about the Centre for Climate Change Economics and Policy can be
found at: http://www.cccep.ac.uk.
The Grantham Research Institute on Climate Change and the Environment was
established by the London School of Economics and Political Science in 2008 to
bring together international expertise on economics, finance, geography, the
environment, international development and political economy to create a worldleading centre for policy-relevant research and training. The Institute is funded by the
Grantham Foundation for the Protection of the Environment and the Global Green
Growth Institute. It has nine research programmes:
1. Adaptation and development
2. Carbon trading and finance
3. Ecosystems, resources and the natural environment
4. Energy, technology and trade
5. Future generations and social justice
6. Growth and the economy
7. International environmental negotiations
8. Modelling and decision making
9. Private sector adaptation, risk and insurance
More information about the Grantham Research Institute on Climate Change and the
Environment can be found at: http://www.lse.ac.uk/grantham.
This working paper is intended to stimulate discussion within the research community
and among users of research, and its content may have been submitted for
publication in academic journals. It has been reviewed by at least one internal referee
before publication. The views expressed in this paper represent those of the
author(s) and do not necessarily represent those of the host institutions or funders.
The Price vs Quantity debate: climate policy and the role of
business cycles1
Anna Grodecka
University of Bonn
Karlygash Kuralbayeva
Grantham Research Institute (LSE), OxCarre and CFM (LSE)
January 24, 2015
Abstract
What is the optimal instrument design and choice for a regulator attempting to control emissions
by private agents in face of uncertainty arising from business cycles? In applying Weitzman’s result
[Prices vs. quantities, Review of Economic Studies, 41 (1974), 477-491] to the problem of greenhouse gas
emissions, the price-quantity literature has shown that, under uncertainty about abatement costs, price
instruments (carbon taxes) are preferred to quantity restrictions (caps on emission), since the damages
from climate change are relatively flat. On the other hand, another recent piece of academic literature
has highlighted the importance of adjusting carbon taxes to business cycle fluctuations in a procyclical
manner. In this paper, we analyze the optimal design and the relative performance of price versus quantity
instruments in the face of uncertainty stemming from business cycles. Our theoretical framework is a
general equilibrium real business cycle model with a climate change externality and distortionary fiscal
policy. First, we find that in an infinitely flexible control environment, the carbon tax fluctuates very
little and is approximately constant, whilst emissions fluctuate a great deal in response to a productivity
shock. Second, we find that a fixed price instrument is advantageous over a fixed quantity instrument due
to the cyclical behavior of abatement costs, which tend to increase during expansions and decline during
economic downturns. Our results suggest that the carbon tax is approximately constant over business
cycles due to “flat” damages in the short-run and thus procyclical behavior as suggested by other studies
cannot be justified merely on the grounds of targeting the climate externality.
Keywords: carbon tax, cap-and-trade, business cycles, distortionary taxes, climate change
JEL Classifications: E32, H23, Q54, Q58
1 We
are grateful to Patrick Kehoe, Per Krusell, William Pizer, Luca Taschini, Rick van der Ploeg, Tony Venables and Martin
Weitzman for valuable comments and suggestions at different stages of the project. We have also benefited from comments by
Simon Dietz, Lawrence Goulder, Garth Heutel, Ethan Ilzetzki, Baltasar Manzano, Sjak Smulders, Frank Venmans, and Jürgen
von Hagen and seminar participants at LSE, University of Oxford, University of Bonn, and the 6th Atlantic Workshop on Energy
and Environmental Economics. Anna Grodecka acknowledges financial support by the Deutsche Forschungsgemeinschaft (DFG)
through the Bonn Graduate School of Economics (BGSE). Karlygash Kuralbayeva acknowledges financial support by Global
Green Growth Institute (GGGI), the Grantham Foundation for the Protection of the Environment and the UK Economic and
Social Research Council (ESRC) through the Centre for Climate Change Economics and Policy. All potential errors are our
own. Contact: [email protected]
1
Introduction
Two classic alternatives for regulating pollutants are a cap-and-trade or a tax; the former is a quantity
control and the latter is a price control. In the face of uncertainty stemming from unexpected changes in
economic circumstances, what is the optimal instrument design and choice for a regulator attempting to
control emissions by private agents? In this paper, we aim to provide an answer to this question.
The literature that compares the relative performance of price and quantity instruments under uncertainty
started with a seminal contribution of Weitzman (1974), who analyzed the optimal instrument choice under
a static partial equilibrium framework, consisting of a reduced form specification of abatement costs and
benefits from abatement. The important character of his setup is that, a regulator issues either a single
price order (fixed price) or a single quantity order (fixed quantity) before uncertainty is resolved and these
fixed policies result in different expected social welfare outcomes under uncertainty. Specifically, Weitzman
shows that under uncertainty about the abatement costs, the relative slopes of the marginal benefit (damage)
function and the marginal cost functions determine whether one instrument is preferred to another. If the
expected marginal benefit function from reducing emissions is flat relative to the marginal cost of abatement,
then a price control is preferred. If, however, the marginal benefit function is steeper, then a quantity control
is preferred2 .
By applying the static Weitzman’s analysis to a problem of controlling pollution caused by emissions
of greenhouse gases, the literature has extended his framework to a dynamic (but still partial equilibrium)
setting (e.g., Hoel and Karp (2002), Newell and Pizer (2003), Karp and Zhang (2005))3 . This literature
emphasizes that, for stock pollutants, such as greenhouse gases, the total stock of pollution changes little
from one year to another, so that the marginal benefit function is basically flat in the short-run. Applying
the Weitzman’s relative-slopes result, the stock pollutant nature of the CO2 appears to make preference for
prices over quantities.
In this paper, we extend this line of research and analyze the design and compare the relative performance
of price and quantity instruments within a general equilibrium framework in case of specific uncertainty business cycles. This analysis is an important extension of the early studies of the price-quantity argument
because of the following reasons. Given the time profile of business cycles - short-run - the intuition of
the price-quantity studies holds within the business cycle setting, implying the regulator favors a price
instrument over a quantity instrument, over the business cycle. These studies however describe situations
when a regulator issues a single-order instrument, which is fixed for at least some period of time, before he
can review it. Contrary to such an approach, a number of recent papers (Heutel (2012), Lintunen and Vilmi,
2013) consider policy settings in which regulators select instruments after the uncertainty is resolved and
investigate the optimal design of carbon pricing instruments over business cycles. This class of models finds
that the optimal carbon taxes and emissions vary over the business cycle in a procyclical way. In addition,
Heutel (2012) finds that carbon taxes fluctuate more than emissions under a carbon tax policy in response
2 Another
implication of the Weitzman’s result is that benefit uncertainty, unless it is correlated with cost uncertainty, does
not affect the net benefit under both price and quantity controls and thus the optimal choice between carbon taxes and emissions
caps. As a result, the many followers of Weitzman (1974) have mostly focused on uncertainty arising from shocks to abatement
costs of firms, with a key exception of Stavins (1996) who shows that under reasonable conditions, correlation between costs
and benefits can reverse the conclusions drawn on the basis of the relative-slope rule
3 The original Weitzman’s analysis has been also extended to analyze performance of hybrid policies that combine elements of
taxes and cap-and-trade schemes (e.g., Roberts and Spence (1976)) and to indexed-instruments (e.g., Newell and Pizer (2003))
3
to a productivity shock.
These are two classes of studies that consider different types of policies: fixed versus state-contingent
policies and they yield different policy recommendations. The first class, starting from a premise that
complex policies that are state-contingent are hard to implement in practice, suggests choosing a fixed
price instrument to address the climate externality in the short-run. The second class suggests making the
stringency of regulation responsive to economic fluctuations. In the case of price regulation, this implies
the carbon tax is increased during economic expansions and is reduced during recessions. In the case of
quantity based regulation, this implies the cap on emissions is relaxed during booms and is tightened during
recessions (Doda (2014)). Better understanding on how climate policies interact with economic fluctuations
and whether business cycles have any bearing on the relative merits of the price over quantity instruments
in the short-run will help in drawing correct and realistic policy implications. This also suggests that the
question of the price-quantity literature needs to be analyzed within a more realistic setting by means of
the real business cycle model. In this way, it would be also possible to link two strands of the literature and
provide additional insight on the design and operation of carbon pricing mechanisms over the business cycle.
Our theoretical framework is an extension of the model in Heutel (2012) by incorporating distortionary
fiscal policy. We calibrate the model to the US economy and use it to investigate the design and dynamics
of optimal carbon tax and cap-and-trade policies. As in Heutel (2012) optimal climate policies appear to be
contingent policies due to the way uncertainty is modeled within the framework: After the shock is realized,
the regulator chooses the instruments to reflect the new contingency of the state of the economy and to
facilitate the adjustment to the shock. State-contingent policies are ideal instruments and thus serve the
role of benchmarks in our analysis in assessing the relative merits of single-order instruments regulators
typically choose in practice. Single-order or “basic” instruments in our model are either a carbon tax or a
cap on emissions fixed at their corresponding steady-state values. We perform welfare analysis and compare
welfare losses from implementing basic regulatory policies instead of baseline policies. In that way, we draw
comparisons with the existing price versus quantity literature and analyze which instrument is preferred
under business cycles, if regulators issue a single order instrument (as in reality). In our model, uncertainty
arises from productivity shocks which indirectly affect abatement costs by firms; the model can also include
the shocks directly to abatement costs. The general equilibrium nature of our theoretical framework allows
modeling and calibrating these shocks explicitly. This is an advantage of our model over the early studies of
the price-quantity literature that have analyzed shocks to abatement costs within a reduced form specification
(typically quadratic) of costs and benefits of pollution.
Simulations of the model produce several results. First, we demonstrate that in an idealized world,
in which planners can continually adjust instruments to reflect current contingencies of state of economy,
considerations about uncertainty in economic shocks are irrelevant. Specifically, the expected welfare outcome
and stock of pollution of greenhouse gases is the same irrespective whether a regulator uses a baseline price
or a quantity instrument.
Second, under such idealized conditions, we show that if a regulator chooses a carbon tax as an instrument,
it remains approximately constant over the business cycle. If a quantity restriction is chosen, then the optimal
quantity varies over the business cycle. These results are explained as follows. A carbon tax fixes the price
of emissions (price of carbon) with an equilibrium quantity of emissions determined by the market. In
contrast, the cap-and-trade fixes the quantity of emissions and leaves it to the market to determine the
price of permits, or shadow price of a unit of emissions. Following Pigou’s principle, the private sector’s
4
cost of emissions (i.e. the carbon tax in case of price instruments and the price of permits, in the case
of a cap-and-trade) must correspond to the marginal damages of pollution. Thus, the optimal carbon tax
must be essentially constant during that period according to the idea (the same as studies in the price
versus quantity literature) that the damages from climate change are essentially constant in the short-run.
In the case of a quantity based policy, in each period, before uncertainty is resolved, the level of emissions
mandated by the regulator deduces the price of permits by the marginal abatement cost meeting the emissions
constraint. In our analysis, uncertainty comes from business cycles which drive the marginal costs. During
a period of economic expansion the cost of abatement tends to increase as more firms conduct abatement,
whereas recessions reduce the demand for polluting goods, which in turn tend to lower the expected costs
of abatement. With the price of permits corresponding to the (approximately) constant level of damages in
the short-run, and with marginal costs varying over business cycles, the optimal quantity restriction must
vary as well.
Finally, we observe that if a regulator chooses either a fixed price instrument or fixed quantity instrument,
then taxes are a more efficient instrument than a quantity instrument. Intuitively, if under idealized conditions, as discussed above, the carbon tax is approximately constant and emissions vary more than welfare
costs associated with fixing the instrument at its steady-state value. This should generate smaller losses
under the former than under the latter policy. Our estimates of such welfare losses confirm this intuition:
we find a fixed tax policy leads to a welfare loss of USD 232.83 per capita per annum, as opposed to the
fixed quantity instrument that generates a loss of USD 247.31 per capita per annum. We also note that even
though both instruments are fixed, firms as well as the rest of the economy can continually adjust to the
shocks. For instance, by comparing impulse responses under both baseline and fixed quantity policies, we
show that pronounced differences in the responses of the variables under these two policies appear only at the
firm’s level and specifically in abatement spending and respectively in the fraction of emissions abated. All
other variables, and specifically consumption of private and public goods exhibit almost identical responses,
suggesting that the welfare costs of fixing quantity are not significantly higher than the welfare costs that
involve fixed price policies.
Some of our analysis assumed state-contingent policies under which a regulator can continually readjust
the instruments. It is apparent that, in reality, it is not feasible for a regulator to quantify shocks affecting
the economy and to continually readjust instruments to reflect changes in economic conditions. This means
that whether a regulator would choose fixed price or fixed quantity instrument will matter for the welfare
outcome. Our results have important implications for a regulator selecting a fixed price or a fixed quantity
instrument in the face of uncertainty coming from unexpected changes in economic circumstances. We argue
that a price instrument is preferred to a quantity instrument due to the cyclical behavior of abatement
costs that fluctuate with business cycles. The dynamics of abatement costs make price controls superior
for controlling emissions when faced with unexpected fluctuations in economic conditions. In reality, if a
regulator chooses a price instrument, then he has to estimate only the level of marginal damages to guide
the level of carbon tax. In the short-run, the level of damages is constant. If, however, he decides to choose a
quantity instrument, then within the context of our framework, it means that the planner has to re-estimate
the marginal costs of abatement by firms every period to make an optimal choice on quantity restriction. Such
one-dimensional uncertainty associated with setting carbon tax and two-dimensional uncertainty associated
with setting the quantity restriction favors the former over the latter. In addition, in practice, regulators
5
likely face an information gap on their side4 : firms likely possess better information about abatement costs
than the regulator because they are closer to the actual production process. This further reinforces our
argument in favor of prices.
Our main finding that the carbon tax is almost constant and fluctuates less than emissions in response
to a TFP shock contrasts the procyclicality and higher volatility of carbon taxes than emissions found by
Heutel (2012). In this paper, we make progress in understanding this fundamental difference in our results
and argue that carbon taxes play a macroeconomic stabilization role as well as correcting the climate change
externality in Heutel’s model. Drawing the parallel with the finding of the optimal commodity taxation
theory that tells us that energy taxation is unlikely to be justified merely on the grounds of raising public
revenues and targeting externalities, we argue that it would be less appropriate to use carbon taxes over and
above that needed to correct externalities over the business cycle.
The rest of the paper is organized as follows. Section 2 describes the core model. Section 3 discusses
the model’s calibration. Section 4 presents and discusses the results under price and quantity instruments
when the regulator can continuously adjust instruments to reflect current states of nature. The same section
presents results when the economy is hit by two correlated shocks - productivity and shock to abatement
technology; it also discusses policy implications of our main results. Section 5 discusses what drives divergence
in our results when compared to those of Heutel (2012). Section 6 concludes.
2
Real business cycle model with distortionary taxes and climate
externality
The baseline model used in this paper is an extension of the real business cycle model with climate externalities of Heutel (2012) by introducing distortionary fiscal policy. The economy consists of households, firms,
and the government. Households obtain utility from consumption of both public and private goods, as well
as from leisure. Goods are produced using private capital and labor. Following Heutel (2012), production
causes greenhouse gas emissions, which accumulate in the atmosphere and lead to climate change that causes
damages by reducing output according to a damage function. As in Heutel (2012), we assume that firms
can counteract the adverse productivity effect of climate change by increasing spending on abatement. The
government levies emissions, corporate and labor taxes on firms. The raised revenue from these taxes are
used to finance the public good provision and the public debt.
2.1
Households
A representative household maximizes:
U = E0
∞
X
β t u(ct , lt , gt )
(1)
t=0
In this utility function ct and gt represent private and public consumption, lt is the number of hours worked
by the household. The representative household faces the following budget constraint:
ct + it + ρBt bt+1 = wt lt + rt kt−1 + πt + bt
4 As
in the original Weitzman’s analysis and in his many extensions
(2)
6
where it is private investment, πt is firm profits, bt+1 denotes one-period government bond purchases, ρBt is
the price of one-period bonds. Households derive income by supplying labor and capital to firms at rental
rates wt and rt . The private capital stock is accumulated according to:
kt = (1 − δ)kt−1 + it
(3)
First-order conditions of the household maximization problem imply:
wt = −
u0L (t)
u0c (t)
(4)
u0c (t) = βEt u0c (t + 1)[1 − δ + rt+1 ]
(5)
u0c (t)ρBt = βEt u0c (t + 1)
(6)
Equation (4) equates the marginal rate of substitution of leisure for consumption to real wages and defines
household’s labor supply. Condition (5) is a standard stochastic Euler equation, which determines intertemporal allocation: it equates the intertemporal marginal rate of substitution in consumption to the real rate
of return on private capital. Condition (6) is the counterpart of equation (5) for domestic bonds.
2.2
Final goods production
Output yt is produced by identical firms, and then can be used for consumption, investment, abatement or
government spending:
yt = (1 − d(xt ))f (kt−1 , lt ; at )
(7)
where at represents an exogenous productivity shock that follows a stationary stochastic process:
ln at = ρ ln at−1 + εt , εt ∼ N (0, σε2 ), | ρ |< 1,
(8)
We assume that the stock of pollution in the atmosphere, denoted by xt , adversely affects output through
damage function d(xt ) specified in the parametrization section 3. The formulation of climate damages as
a fraction of output lost as in (7) was introduced by Nordhaus (1991) and since then has been extensively
used in the literature. The mapping of emissions to economic damage can be thought as comprising of two
steps: first, emissions increase the concentration of greenhouse gases leading to climate change (represented
by change in the global mean temperature), and second, changes in temperature cause economic damages.
Some papers, e.g., Barrage (2014) follow Nordhaus’s approach and model two steps of mapping from carbon
concentration to damages. We follow equally common specification and map CO2 concentration to damages
in one step (Heutel (2012), Golosov et al. (2014))5 . The specification (7) assumes that climate change (or
concentration of greenhouse gas emissions in our set up) affects output directly. Such specification is standard
5 The
two stage mapping would have introduced a set of lags in the effect of current-emissions on output, but would have not
changed our results. This is because cyclical changes in emissions levels have very little effect on the pollution stock because of
the long-lived nature of CO2, and thus it is relatively immaterial whether a ton of carbon dioxide emitted today or a few periods
later. We demonstrate in the Online appendix that the damages from pollution do not change significantly with business cycles.
7
in climate change modeling analysis and it represents in aggregate form the dependence of production of
many goods on climate change conditions, such as production of agricultural goods, forestry, fisheries etc.
Profits of firms are defined as:
πt = (1 − τct )yt − wt (1 + τLt )lt − τEt et − rt kt−1 − zt
(9)
where τLt is payroll (labor) tax, τct is corporate tax, τEt is tax on emissions, et are emissions, which are
by-product of production, and zt is spending on abatement by firms; private abatement spending is assumed
to abate the µt fraction of emissions via the following relation:
zt
= m(µt )
yt
(10)
so that firms face the emissions constraint given by:
et = (1 − µt )h(yt )
(11)
where h(yt ) determines total emissions from producing yt output. Following Heutel (2012), we assume that
a climate change externality arises because firms do not take into account their emission’s impact on the
pollution stock and thus on productivity. In other words, firms take xt as a given. Optimality conditions of
the firm imply:
rt = (1 − d(xt ))fk0 [1 − τct − τEt (1 − µt )h0 (yt ) − m(µt )]
(12)
wt (1 + τLt ) = (1 − d(xt ))fL0 [1 − τct − τEt (1 − µt )h0 (yt ) − m(µt )]
(13)
τEt =
yt m0 (µt )
h(yt )
(14)
Equation (12) is an optimal condition of demand for capital, which implies that the return associated with
an increase in capital stock by one unit is equal to the marginal product of capital, net of additional tax
payments on increased emissions associated with an increase in output and net of additional spending on
abatement to clean a given fraction µ of extra emissions stemming from an increase in output. Equation
(13) is the counterpart of equation (12) for labor demand. Finally, equation (14) says that the firm reacts
to the carbon tax by choosing the level of abatement (equivalently the level of emissions) such that the tax
on emissions would be equal to the marginal cost of emissions reduction.
2.3
Government
The government budget constraint is balanced according to:
gt + bt = wt τLt lt + τEt et + τct yt + ρBt bt+1
(15)
where the government raises revenues by taxing labor income and emissions and levying corporate tax to
finance public debt bt and provision of public goods, gt . The government can issue new one-period bonds
bt+1 . The government budget constraint (15) incorporates market clearing for bonds which requires that
households demand for bonds and government supply for bonds are equated.
8
2.4
Carbon cycle
Following Heutel (2012), we assume that each period new industrial domestic and foreign carbon dioxide
emissions increase the existing pollution stock that decays at a linear rate η:
xt = ηxt−1 + et + erow
t
(16)
where et is current-period domestic emissions that are related to the output produced and fraction µt that is
abated, while erow
is current-period emissions from the rest of the world and η is the fraction of the pollution
t
stock that remains in the atmosphere.
Atmosphere, however, is not the only reservoir of the carbon dioxide. Even without industrial emissions
there exists a natural carbon cycle encompassing flows of carbon dioxide among different reservoirs, such as
atmosphere, oceans etc. Nordhaus (2008) distinguishes in his model between three of them: the atmosphere,
biosphere including the upper oceans, and the deep oceans. The flows between the first two reservoirs are
relatively quick, but the carbon dioxide from deep oceans, that is the largest carbon reservoir, interacts only
at a very slow pace with the remaining two reservoirs. In Nordhaus (2008) only the atmosphere is directly
influenced by industrial activity. The mass of carbon accumulated in the atmosphere has in turn impact
on the radiative forcing that enters the damage function adversely affecting the production process. The
carbon accumulated in each of the reservoirs is a function of past carbon values, depreciated by a parameter
analogous to our decay rate η. So, in this paper, we leave the flows between atmosphere and other carbon
reservoirs aside and model the effect of industrial activity on atmosphere. One-dimensional representation
of the carbon cycle based on the stock of pollution in atmosphere only has been also utilized in Golosov et
al. (2014).
2.5
Characterizing equilibrium
To construct the Ramsey problem, we reorganize some of the constraints in order to reduce the number
of choice variables and to obtain a compact expression for the household budget constraint. In particular,
combining (2),(9) and (15) gives the following resource constraint for the economy:
ct + kt − (1 − δ)kt−1 + zt + gt = yt
(17)
Next, by adding and substituting for wt from (4), we rewrite the government’s budget constraint as follows:
gt + bt = −
u0L (t)
τLt lt + τEt (1 − µt )h(yt ) + τct yt + ρBt bt+1
u0c (t)
(18)
Substituting (4) into (13) gives:
−
2.6
u0L (t)
(1 + τLt ) = (1 − d(xt ))fL0 [1 − τct − τEt (1 − µt )h0 (yt ) − m(µt )]
u0c (t)
(19)
Ramsey problem
Ramsey planner maximizes the utility of households:
E0
∞
X
t=0
β t u(ct , lt , gt )
(20)
9
subject to (5), (14), (17), (18), (19) and
yt = (1 − d(xt ))f (kt−1 , lt ; at )
(21)
xt = ηxt−1 + (1 − µt )h(yt ) + erow
t
(22)
where we also use the function (12) for definition of rt .
The government chooses ct , µt , kt , yt , xt , lt , τLt , τEt , τct , gt and bt+1 to maximize (3) subject to the
constraints specified above.
The Lagrangian for this problem is given by:
L =
E0
∞
X
β t {u(t) + λt [−u0c (t) + βu0c (t + 1)(1 − δ + rt+1 )] +
t=0
+Ωt [ct + kt − (1 − δ)kt−1 + m(µt )yt + gt − yt ] +
+χt [τEt h(yt ) − yt m0 (µt )] +
u0 (t)
τLt lt + τEt (1 − µt )h(yt ) + τct yt + ρBt bt+1 ]
+Λt [−gt − bt − L0
uc (t)
+λpt [yt − (1 − d(xt ))f (kt−1 , lt , kGt−1 )] +
0
u (t)
+ςt − L0
(1 + τLt ) − (1 − d(xt ))fL0 (t)(1 − τct − τEt (1 − µt )h0 (yt ) − m(µt )) +
uc (t)
+Φt [xt − ηxt−1 − erow
− (1 − µt )h(yt )]
t
The first-order conditions of the Ramsey problem are given in Appendix 8.3.
3
Parametrization
In calibrating the model, we select parameter values that enable the theoretical model to generate features
that are (as closely as possible) consistent with the main features of the US economy. We assign values to
structural parameters using values that are common in business cycle studies of fiscal policy and macroeconomic models with climate change externalities. In calibrating the climate part of the model, we draw
strongly on estimates and parameter values used in Heutel (2012). Baseline parameter values of the model
are summarized in Table 1, while Table 2 reports macroeconomic ratios implied by the theoretical model as
well as the corresponding values for the US data. Data sources employed in these calculations are summarized
in Appendix 8.2.
In calibrating the model, a time period represents one quarter. The production function is given by
α
f (kt−1 , lt ; at ) = at kt−1
lt1−α
(23)
We set α at 0.36, which is a value commonly used in the standard RBC literature. For the TFP process, we
assume that ρ = 0.95 and σε = 0.007, where the value of the standard deviation is as in Heutel (2012) and
is similar to the value 0.0056 as in Schmitt-Grohe and Uribe (2007). We show below that our results are not
sensitive to changes in the value of ρ. The private capital depreciation rate, δ, is set at 0.025 (Heutel, 2012).
We set the discount factor β at 0.98.
For the quantitative analysis, we consider the following form of the households utility function:
u(ct , lt , gt ) =
c1−κ
−1
g 1−κ − 1
l1+ψ
t
+θ t
− t
1−κ
1−κ
1+ψ
(24)
10
Parameter
Value
Definition
α
0.36
private capital share in the production function
ρ
0.95
persistence of the TFP shock
σε
0.007
standard deviation of the TFP shock
δ
0.025
private capital depreciation rate (quarterly)
β
0.98
subjective discount factor (quarterly)
κ
1.6
coefficient of relative risk aversion
θ
0.236
weight of public consumption in utility
1/ψ
0.4
Frisch elasticity of labor supply
η
0.9979
pollution decay
d2
5.2096e-10
damage function parameter
d1
-1.2583e-06
damage function parameter
d0
1.3950e-3
damage function parameter
θ1
0.05607
abatement cost equation parameter
θ2
2.8
abatement cost equation parameter
1−ν
0.696
elasticity of emissions with respect to output
Table 1: Baseline parameter values
with the coefficient of relative risk aversion, κ, set to 1.6, which implies that the value of the intertemporal
elasticity of substitution (EIS) is 0.625 in the model. The standard value of κ in the literature is 1 (see, e.g.,
Golosov et al., 2013). We set the value of ψ such that a Frisch elasticity of labor supply is 0.4, in line with
macroeconomic estimates reported by Rogerson and Wallenius (2009). The weight of public consumption in
the utility function, θ is set at 0.236.
Following Heutel (2012), the pollution stock in the atmosphere evolves according to the following equation:
. We set the value of η at 0.9979 as in Heutel (2012), who calibrated this parameter
xt = ηxt−1 + et + erow
t
assuming that 83 years represent the half-life of atmospheric carbon dioxide. In actuality, there is no
single number that describes the lifetime of carbon dioxide in the atmosphere because it is weighted sum of
exponential decays at different rates. Carbon dioxide is not destroyed in the air, but is instead exchanged
between the atmosphere, the ocean, and land. For other greenhouse gases, lifecycle estimation is possible,
see the report of Intergovernmental Panel on Climate Change (2001). Thus, the values used by different
studies vary. Following Archer (2005), Golosov et al. (2014) we calibrate the half-life of CO2 to 150 years.
Assuming a half-life of 150 years would lead to η = 0.9988 in our model. Our model’s result are not sensitive
to changes in η - the decay parameter influences quantitative responses of only three variables (emissions,
stock of pollution and fraction of emissions abated), while the responses of all other variables remain the
same.
The emissions produced by the rest of the world, erow
are set to 4 times the steady state of domestic
t
emissions, which is guided by the following considerations. According to data by the U.S. Environmental
Protection Agency (EPA), the USA accounted for 19% of global CO2 emissions from fossil fuel combustion
in 2008, which means that global emissions were four times higher than those in the US6 .
6 http://www.epa.gov/climatechange/ghgemissions/global.html
11
The loss of potential output due to pollution is governed by the function d(x) = d2 x2 + d1 x + d0 . We
set the values of d2 , d1 , d0 respectively to 5.2096e-10, -1.2583e-06, 1.3950e-3, following Heutel (2012), who
calibrates these values to match the damages from carbon dioxide in the atmosphere estimated by papers
in the environmental literature. Specifically, Heutel (2012) bases this estimation on Nordhaus (2008). The
DICE model includes a damage function, expressing climate damages as a fraction of world output. In the
DICE model, the damages are calculated for the whole world, but the supplementary material7 explains how
the impact of each larger region in the world was taken into account, so the percentage impact of the US
in the world damages is given. Nordhaus’s model (Nordhaus, 2008) contains equations linking the pollution
stock to its radioactive force and its impact on the temperature of oceans and the atmosphere. Having the
pollution stock, one can thus compute damages as a fraction of output, using the model’s equation. Heutel
(2012) does this exercise for 100 different pollution stocks ranging from 600GtC to 1200GtC and plots the
pollution stock against damages on a graph. The resulting functional relationship is fitted to a quadratic
function, which is the d(x) function in the model8 . Our baseline calibration gives damages of 0.59%9 .
The abatement cost function is taken directly from Nordhaus (2008) and has the form m(µ) = θ1 µθ2 .
We set θ1 = 0.05607 and θ2 = 2.8, following Heutel (2012).
Symbol
Variable
Model
Data
c/y
personal consumption/output
0.58
0.68
g/y
government consumption/output
0.26
0.15
i/y
private domestic investment/output
0.16
0.16
e/y
emissions/output
0.76
0.60
b/y
public debt/output
0.77
0.77
µ
fraction of emissions abated, %
0.54
1.85
τE
tax on emissions, %
0.002
-
τL
labor tax, %
15.4
15.4
τC
corporate tax, %
21.25
35
τE e/y
revenue from carbon tax, % of GDP
0.0013
0.7 (estimate)
Table 2: Structure of the theoretical economy and the data
Given our baseline parametrization, the theoretical model implies very low level of carbon taxes in the
steady-state, 0.002% (for comparison Heutel’s model implies 0.0487%), and respectively the very low share
of carbon tax revenues in GDP, only 0.0013% of GDP. We think this is not problematic and does not drive
our results, since the estimates suggest that in reality the share of carbon tax revenues in GDP is a negligible
number anyway. Specifically, different estimates for the US evaluate the possible net revenue in the range
0.51-0.8% of US GDP (Table 2 in Gale et al., 2013). As Gale et al. (2013) report, in 2007 the carbon
tax raised revenues equivalent to 0.3% of GDP in Finland and Denmark and 0.8% of GDP in Sweden. In
7 http://www.econ.yale.edu/
8 It
~nordhaus/homepage/Accom_Notes_100507.pdf
is important to note that the damage function was calibrated using the point estimates of the equilibrium climate
sensitivity, which is highly uncertain with a “likely” range between 1.5 and 4.5◦ C (Intergovernmental Panel on Climate Change
(2013)). In fact, investigations started with analysis in Weitzman (2009) suggests that the climate sensitivity parameter is
better thought as the distribution with fat tails.
9 Given the small values of the parameter values in the damage function and for the relevant values of concentration of stock
of pollution, the damage function would not give damages greater than 100%.
12
Australia carbon tax revenue 2012-2013 accounted to 1.2% of GDP10 .
Finally, output is mapped into emissions through h(yt ) = y 1−ν , with et = (1 − µt )h(yt ), where 1 − ν
represents the elasticity of emissions with respect to output. We set the value of 1 − ν at 0.696, which is
the estimate from a (seasonally adjusted) ARIMA regression of the log of emissions of CO2 on the log of
GDP for US data in years 1981-2003. As in Heutel (2012), we solve the model by log-linearizing around the
steady-state.
4
Simulation results
4.1
Results under baseline price instrument policy
Figure 1 shows the impulse responses (IR) of the key variables to a 1% increase in productivity under both
carbon tax and cap-and-trade policies. All variables are expressed in terms of percentage deviations from
the steady state, except for the tax rates, for which responses are expressed as absolute deviations from
their steady-state values11 . Given the objectives of the paper, we report plots of the impulse response
functions only for key variables related to our analysis, but results for the remaining variables are available
upon request. The continuous line represents the baseline model, the dashed line represents the model with
alternative tax policy. We start with discussing the results under baseline carbon tax policy.
Impulse responses obtained from simulations of the baseline model result in the following key qualitative
results. First, consistent with the findings of other studies on optimal carbon tax over the business cycle (e.g.,
Heutel, 2012), emissions increase in the periods following a positive productivity shock. Given the long-lived
nature of carbon dioxide, increased emissions result in a higher pollution stock over the medium term, and
increase by around 0.008% in 25 years time. We demonstrate in Online appendix 8.1 that this number is
quite small, by estimating corresponding increases in mean global temperature and sea levels associated
with this rise in atmospheric greenhouse gases. Second, the labor tax increases by 1.66 percentage points
(10.77 percent), the corporate tax decreases by 1.22 percentage points (5.75 percent) while tax on emissions
is raised by only 0.000014 percentage points, corresponding to 0.7 percent relative to the steady-state value
in response to the shock.
Very small fluctuations in carbon tax can be explained drawing on the intuition of the “price versus
quantity” literature. Given the long-lived nature of greenhouse gases, the additional damage from each
additional ton of carbon emissions is constant in the short-run. In terms of the model presented in section
2, concentration of CO2 emissions in the atmosphere xt = ηxt−1 + et + erpw
, xt ∼ xt−1 as well as damages
t
d(xt ) = d2 x2 + d1 x + d0 remain essentially constant over the business cycle. Following Pigou’s principle,
the private sector’s marginal cost - carbon tax under baseline policy - must correspond to the level of the
marginal damages, which are “flat” in the short-run. This explains why the optimal carbon tax is essentially
constant over the business cycles12 .
10 http://www.treasury.gov.au
11 Please
note that for each time period t, we plot the values of those stock variables which enter current production process,
namely xt and kt−1 . Since et affects xt contemporaneously, xt jumps in response to the shock, while kt−1 does not.
12 Abstracting from business cycles, Golosov et al. (2014) propose a tractable Ramsey growth model and show that an optimal
carbon tax is proportional to output. Rezai and van der Ploeg (2014) generalize their result to allow for some elements such
as population growth, a temperature lag and general degrees of intergenerational inequality aversion and show that the global
carbon tax rises in proportion with GDP if marginal climate damages are proportional to GDP.
13
4.2
Results under baseline quantity instrument policy
The baseline model assumes that tax on emissions is an instrument to combat climate change. An alternative
policy to control emissions is a cap-and-trade or emissions trading scheme, in which governments restrict the
emissions (by imposing cap on emissions) firms produce. Following Heutel (2012), we introduce a cap-andtrade scheme into our framework, by assuming that the government mandates the level of emissions a firm
can produce, qt . In other words, the government allocates permits to each firm (one representative firm) for
free, so that it does not generate revenue. The setting features the simplest cap-and-trade scheme that does
not allow for policies similar to a “safety valve”, in which firms are allowed to purchase an unlimited number
of permits at a set price, which equivalently sets a ceiling on the price of permits (see e.g., Pizer (2002)
for more details); we also abstract from incorporating active banking, which allows regulated firms to shift
obligations across time in response to periods of unexpectedly high or low marginal costs (see, e.g., Fell et
al. (2012))13 . And since the theoretical framework features one representative firm, the quantity constraint
is equivalent to a cap-and-trade scheme.
An individual’s budget constraint and FOC in this setting remain as in the baseline model. There are
only changes in the firm’s problem and in the government’s budget constraint. Specifically, firms do not pay
taxes and respectively the government budget omits revenues from taxing emissions. Profits of the firms are
defined as:
πt = (1 − τct )yt − wt (1 + τLt )lt − rt kt−1 − zt
(25)
subject to emissions constraint qt = (1 − µt )h(yt ) and abatement spending zt = m(µt )yt . The government
budget constraint is balanced according to:
gt + bt = wt τLt lt + τct yt + ρBt bt+1
(26)
Optimality conditions of the firm imply:
rt = (1 − d(xt ))fk0 [1 − τct − m(µt ) −
m0 (µ)yt
(1 − µt )h0 (y)]
h(yt )
wt (1 + τLt ) = (1 − d(xt ))fL0 [1 − τct − m(µt ) −
m0 (µ)y
(1 − µt )h0 (y)]
h(y)
qt = (1 − µt )h(yt )
(27)
(28)
(29)
Equation (29) is just a constraint on the quantity of emissions produced. Equations (27)-(28) are analogous
to the equations (12)-(13) under tax policy, and they are optimal conditions of demand for capital and labor,
respectively. They also demonstrate that the price of permits - the shadow price of a unit of emissions under
quantity policy - is pEt ≡ m0 (µ)yt /h(yt ). For comparison, under price instrument, as shown in equation
(14), firms reduce emissions until the marginal cost of reductions equal to the tax - price of carbon. In other
words, carbon tax fixes the price of emissions, so that the equilibrium quantity is determined in the market; in
contrast, the cap-and-trade fixes the quantity of emissions and leaves it to the market to determine the price.
In a deterministic world, the carbon tax under priced policy would be equal to the shadow price induced from
13 Active
banking can make cap-and-trade scheme more flexible in terms of intertemporal allocation of abatement decisions
by firms. As a result, in face of temporary uncertainty in costs, under cap-and-trade with banking and borrowing, emissions
fluctuate period-by-period and prices are relatively constant (Parsons and Taschini (2013)).
14
cap on emissions and two instruments lead to the same emissions outcome. In such world, there is simple
equivalence between policies: a given price yields a specific quantity of emissions and vice-versa. Under
uncertainty, however, and if policies must be fixed before the uncertainty is resolved as in the framework of
the price-quantity literature, two policies lead to different outcomes (see, e.g., Weitzman (1974)), and the
price of carbon under priced policy would not be equal to the induced price from cap-and-trade.
In our framework, under uncertainty in business cycles driven by the same productivity shock, both
policies lead to the same expected welfare outcome and optimal quantity under cap-and-trade varies with
business cycles. Plots of impulse responses of key variables, 1, demonstrate that both policies lead to the
same expected welfare and emissions outcome. All variables, except tax on labor, wages, government bonds,
spending on abatement and fraction of emissions abated, exhibit identical responses. The above-mentioned
variables respond differently because the government does not generate any revenue from a cap-and-trade
and thus these variables need to adjust accordingly to generate the expected welfare outcome as under price
instrument.
Since in our model regulators can continually readjust instruments to reflect changes in economic circumstances, both lead to the same expected welfare outcome. Another result worth mentioning is that under
cap-and-trade policy, the optimal restriction of emissions varies with business cycles. The intuition for this
result becomes clear and simple after understanding the instrument’s mechanism. In each period before the
uncertainty is resolved, the regulator mandates the level of emissions, which will deduce the shadow price
of carbon by the marginal cost meeting the emissions constraint. In our framework uncertainty comes only
from business cycles, which will then affect the level of marginal costs, which tend to increase during booms
and to fall during recessions. Thus, every period when the uncertainty is resolved, the state of the nature
will be associated with a different marginal cost. Following Pigou’s principle (Pigou (1920)), private sector’s
cost - the shadow price of carbon - must correspond to the marginal damages of pollution. Thus, with
essentially constant level of damages in the short-run, and with varying over the business cycle marginal
cost, the optimal quantity restriction must vary with business cycle to deduce a shadow price that is not
only consistent with the target for emissions, but also internalizes externality.
4.3
Fixed priced and fixed quantity based policies and welfare
As discussed above, when the regulator can continually readjust the policies, the choice of the optimal
instrument - price or quantity - becomes irrelevant as both policies lead to the same expected welfare
outcome. Feasibility of such complex policies can be doubted in the practice and we discuss the implications
of our results for policy analysis in section below. In contrast, the price-quantity literature, initiated with
analysis in Weitzman (1974), focuses on the consequences of “basic” policies, those when the regulator chooses
either a fixed price or fixed quantity policy before any uncertainty is resolved. To follow this convention,
in this section we investigate the relative performance of fixed price and fixed quantity policies (fixed at
corresponding steady-state values), by comparing welfare losses from fixing policies compared to the baseline
policies.
Our measure of welfare is the amount of baseline steady-state policy consumption a household would be
willing to give up to be as well off under the alternative specification as under the baseline policy, following
the procedure of Schmitt-Grohe and Uribe (2007). The results are shown in Table 3. For the consumptionequivalence, a number of, e.g., 0.64 means that the alternative environmental tax policy reduces welfare by
15
0.64% of consumption on average.
Welfare in consumption-equivalents, %
Model with fixed emissions tax
0.64
Model with fixed quantity
0.68
Table 3: Welfare effects of alternative tax policies
We express welfare costs associated with single order instruments in monetary value, using the 2013
US annual personal consumption expenditure14 , which stood at USD 11,496.2 bn. By using this data, and
converting this to per capita terms15 , we find a fixed tax instrument does in fact lead to a lower welfare
loss compared to the fixed quantity instrument: USD 232.83 per person with taxes vs. USD 247.31 under
quantity controls16 .
These relatively small differences in the welfare losses under tax instrument compared to the quantity
instrument can be explained as follows. Even though, both instruments are fixed, firms as well as the rest of
the economy can continually adjust to the shocks. For instance, the impulse responses under both baseline
and fixed quantity policies demonstrate (figures 2 and 3), pronounced differences in the responses of the
variables under these two policies appear only at the firm;s level and specifically in abatement spending
and respectively in the fraction of emissions abated. And as welfare comprises consumption of both private
and public goods, it is not surprising to see small differences in welfare costs under fixed tax and fixed
cap-and-trade are justified. In line with that, we will show in the next section that responses to the shock to
abatement technology also occur primarily at the firm’s level. Such adjustment occurring at the firm’s level
will have implications for policy conducted in reality as we will discuss policy implications of our results in
section 4.5.
Finally, some other studies also find very small differences in the welfare gains from contrasting different
policy instruments, even though those estimates are not directly comparable with ours. In particular, Pizer
(1999) investigates the relative performance of taxes with rate controls (fractional reduction CO2 emissions
at a given time) in an integrated climate-economy model under uncertainty which is modeled allowing
thousands of different states of nature. He finds that uncertainty leads to a preference for taxes over control
rates, with the optimal rate control generating welfare gains17 equivalent to a USD 73 increase in current
per capita consumption, whilst the optimal tax policy generates an USD 86 increase.
14 Data
source is the NIPA table, see Appendix 8.2 for more details.
in the US in 2013 stood at 316.1 million people.
16 The uncertainty in our paper arises from temporary shocks and our results are not sensitive to changes in the persistence
15 Population
of the shocks. See figure 5 in the Online appendix that presents the IRFs under different values of the persistence of the shock
under carbon tax policy. The welfare ranking of the instruments also remain unchanged and the results are available upon
request. But in general, dynamic structure of cost uncertainty can affect the choice between a price or quantity control, as
shown in Parsons and Taschini (2013). Specifically, by using reduced form specification in tradition of the early price-quantity
literature, they show that temporary shocks to abatement cost favor the use of a price control, whilst the permanent shocks
favor a quantity control.
17 The source of such gain is due to those states of the nature in which the marginal costs of reduction of emissions are
low, while the marginal benefits are high, which favor more stringent policies. While opposing states of nature that favor less
stringent policies and thus generate losses from more stringent policies, but such losses are not as significant as the gains,
resulting in overall improvement in welfare. In other words, more stringent than the optimal control rate policy ignoring
uncertainty improves welfare.
16
4.4
Associated shocks to abatement technology
The one of the key underpinnings of the argument we proposed for comparing the relative merits of alternative
price and quantity mechanisms is that price instrument gives flexibility to firms to find their own most efficient
solutions in controlling emissions. To provide further evidence for that, we perform next experiment, in which
we assume that the economy is hit by two, correlated shocks, productivity shock and shock to abatement
technology. Such experiment has been motivated by the following considerations.
In our baseline model, uncertainty comes from the productivity shock. The existing “price versus quantity” literature, however, models a reduced form of the abatement cost function with mean-zero random
shocks to marginal abatement costs. The shocks to the reduced form of abatement costs may originate
(indirectly) from productivity shocks or directly from business cycles. In our framework, we can differentiate
between these two types of shocks to abatement costs, by considering productivity shock and an abatement
shock. We introduce an abatement shock as a shock to abatement technology εab,t :
zt
= m(µt )εab,t
yt
(30)
which assumed follows AR(1) process, defined as:
ln εab,t = ρεab,t ln εab,t−1 + ρab εt ,
(31)
where εt is the shock to productivity, and ρab εt is a shock to abatement technology. Following the discussion
above, we assume ρab > 0. Note that we have defined the shock to abatement technology such that a positive
value of ρab εt increases abatement costs, that is, abatement of a given fraction of emissions µ associated with
a given output becomes more costly. As mentioned earlier, there are two new values in this extension that
we need to parametrize: the value of the persistence of the shock to abatement technology, and the value of
correlation between shocks to productivity and abatement. Since we assume that abatement costs vary with
business cycles, we can set the value of persistence of the shock to abatement technology equal to the one
of productivity shock. And since the value of the correlation between productivity and abatement a priori
is unknown, we experiment with two values of ρab : 0.4 and 0.7.
Comparison of impulse responses under the baseline policy and under correlated shock case (figure 4)
reveals that adjustment to the shock to abatement technology happens through changing the total spending
on abatement, without any notable effects on the behavior of the remaining variables. As a result, the firm
produces the same level of emissions and abate the same fraction of emissions. To sum up, firms find their
own most efficient solutions to controlling emissions.
4.5
Policy implications
State-contingent policies18 considered in the model are difficult, if not impossible, to implement in practice
because they involve continual readjustment of policies and require complete knowledge about distribution
of shocks affecting economy. Despite these arguments, our baseline results provides important policy implications and insights for a policymaker seeking a policy regulation - fixed price or fixed quantity restriction to control CO2 emissions in face of uncertainty stemming from business cycles.
18 Newell
and Pizer (2003) extended the original analysis of Weitzman’s to indexed policies, where quantities are proportional
to an index, such as economic output. They find that a general indexed quantity policy improves the ex post performance
of fixed quantities, but comparison to a fixed price policy is more complex. But they point out that identifying the proper
economic activity indicator is a complex task: the indicator must capture the direction and the right intensity of the shock.
17
Specifically, under priced policy, our results suggest that in practice the regulator has to estimate the
level of marginal damages to inform the level of carbon tax. As seen in the model, firms react to a carbon
tax, by reducing emissions until the marginal cost of abatement equals to the tax. Conversely, if regulator
selects quantity based policy, he must estimate both the level of the marginal damages and the marginal
cost of abatement to deduce the target for emissions, which induce the shadow price of emissions that
internalizes the externality. But, the level of marginal costs vary with business cycles and to be able to set
a target that yields the economically efficient outcome, he must re-estimate marginal costs every period.
Thus, one-dimensional uncertainty associated with setting the carbon price compared with two-dimensional
uncertainty associated with setting quantity target argues in favor of the former over the latter instrument.
In drawing this policy implication, we were referring to a genuine uncertainty stemming from business
cycles and uncertainty in estimating the marginal damages (so-called social cost of carbon) that exist not
only for regulator but also for producers. However, in reality, another type of uncertainty may be present
- information gap - randomness that is certain to a producer but is unknown to a regulator or vice-versa.
Specifically, it is plausible to assume, as in the original Weitzman’s analysis (Weitzman (1974) and as in most
of studies that have followed) that uncertainty in the marginal costs function is an information gap on the
side of the regulator19 . That is, firms possess better information about costs than the regulator because they
are actually closer to the actual production process. The presence in reality of the type of information gap
as described above reinforces our argument in favor of prices. Carbon taxes, helping in controlling emissions,
likely provide firms and businesses with flexibility to innovate and find their most efficient solutions, whilst
not requiring for a regulator to face a difficult task of estimating marginal costs of abatement by firms. As
we discussed, for instance in the previous section, even under the idealized circumstances when regulator can
continually adjust instruments, the adjustment to the shock to abatement technology happens at the firm’s
level.
This reasoning in superiority of price over quantity echoes an argument of Pizer (2003) in favor of price,
but without formal analysis of this paper20 :
Rather than attempting to hit a fixed quantity target at any cost, we should instead price
emissions at our best guess concerning their rate of marginal damage. Since there is a real risk
that the costs of hitting a fixed quantity target can be extremely high - depending on growth
and technology - such targets make little sense.
Finally, we find duality in our argument - one-layer of uncertainty vs two-layer of uncertainty in face of business cycle shocks- with another idea of Weitzman, laid out in his recent paper Weitzman (2014). Weitzman
contrasts the properties of an idealized binding harmonized price with an idealized binding cap-and-trade
system within the context of international negotiations that aim solving global warming externality problem.
He argues that setting an internationally-harmonized carbon price involves only one layer of negotiations as
opposed to two on quantity side. His basic intuition is as follows. Under a quantity-based system, n countries
19 Laffont
(1977) provides a detailed discussion of the information structure present in policy choice problems by regulator
choosing between prices or quantities in tradition of the original Weitzman’s analysis.
20 Pizer (2003) tests the robustness of the claim that under possibility of catastrophic damages, quantity instrument is preferred
instrument. The existence of some thresholds of climate change is one of the few arguments for quantity-based regulations,
as most analyses, including the current paper, overwhelmingly argue in favor of price rather than quantity instrument. Other
arguments in favor of quantity may include political economy and many administrative challenges and for comprehensive
discussion of the advantages of carbon tax vs cap-and-trade, see, e.g., Hepburn (2006) and Goulder and Schein (2013).
18
participating in negotiations must agree on the single aggregate level of emissions and on the distribution of
aggregate emissions among n parties. By contrast, a price-based system of negotiations focus on agreeing to
a single one-dimensional uniform price.
5
Carbon taxes and business cycles
We have shown that carbon tax is approximately constant over the business cycle and our results are extension
of the findings of the “price versus quantity” literature to a general equilibrium framework. Impulse response
function results also demonstrate that emissions exhibit larger volatility than taxes. But our results are in
contrast with the findings of the Heutel (2012) who points out to procyclical behavior of carbon taxes in
response to business cycle shocks. Moreover, he finds that carbon taxes fluctuate by more than emissions in
response to a productivity shock. In this section we attempt to understand what drives divergence in our
results.
The procyclicality result of carbon taxes in the Heutel’s model can be explained, by referring to the
optimal conditions of the firms and household’s Euler equations:
rt = (1 − d(xt ))fk0 [1 − τEt (1 − µt )h0 (yt ) − m(µt )]
τEt =
yt m0 (µt )
h(yt )
uc,t = βEt uc,t+1 [1 − δ + rt ]
(32)
(33)
(34)
The equation (33) is identical to the equation in our model (14) and represents the role of carbon taxes
internalizing the climate externality. The setting of the theoretical framework in the Heutel’s model however
implies that carbon taxes also distort capital accumulation and thus return on capital (32) and thus affect
intertemporal reallocation of consumption, through Euler equation (34). This means that Ramsey planner
uses carbon taxes to facilitate consumption smoothing across periods. Intuitively, as abatement is costlier
during economic expansions, the carbon tax must to rise to prompt firms avoid producing more emissions
during expansions; opposite is true during declines in economic activity. This facilitates intertemporal reallocation of emissions across periods: emit less today than otherwise during boom but be compensated for
that with higher than otherwise emissions during recessions. As emissions are by-product of output, such
trade-off in emissions creates intertemporal reallocation of consumption. In line with this, Heutel points
out that: ”It is variance in consumption, not in pollution stock, that leads to the variance in the emissions
tax”. In such way, carbon taxes end up playing role that it is initially not subscribed to, and in particular
macroeconomic stabilization role. Such “non-standard” outcome usually appears in the optimal taxation
literature when the tax system is incomplete. The tax system in the Heutel’s model is indeed incomplete
in the sense that there are more competitive equilibrium conditions in which taxes are involved than tax
instruments (Chari and Kehoe (1998)). These equations are (32) and (33).
Completeness of the tax system is important for at least two reasons. First, as shown by Chari and Kehoe
(1998), Correia (1996), Aruoba and Chugh (2010) and many others, an incomplete tax system requires that
new constraints reflecting this incompleteness to be added to the Ramsey problem. Second, incomplete
tax systems can lead to“non-standard” policy prescriptions because some instruments end up serving as
19
imperfect proxies for other, unavailable instruments.21 . Thus to ensure completeness of the tax system in
the framework similar to one in Heutel, there is need of introducing one additional distortionary tax, but
since we also incorporate labor into his original model, we need to introduce two distortionary taxes: on
labor and corporate income, and under such setting carbon tax would only play the role it is introduced
originally for - correction of climate externality.
The conclusion that emerges out of the above discussion has important policy implications. It suggests
that taxation of emissions cannot be justified on the grounds of macroeconomic stabilization tool and other
than to target climate change externalities. This is similar logic to the conclusion of the optimal taxation
theory applied to the taxation of energy and energy related products, that pure revenue raising is best done
with wide-base taxes, such as VAT or taxes on labor, rather than carbon taxes22 .
6
Conclusion
The relevance and importance of the analysis of an optimal policy instrument for a regulator seeking to
control CO2 emissions in the face of unexpected fluctuations in economic activity has increased very recently,
particularly in the aftermath of the financial and economic crisis of 2008. In the wake of the global financial
crisis, knowing whether environmental policies should be accommodative to unexpected changes in economic
conditions has received intense interest. Such interest has been partly prompted by a marked and persistent
drop in the price of permits within the largest cap-and-trade system, the EU’s Emissions Trading Scheme (EU
ETS). As argued by many observers, this was mainly driven by a combination of low demand for emissions
permits caused by the recession and inflexibility of the caps on emissions to changes in economic conditions.
The debate is under way on how the EU ETS system needs to be reformed to make the system more resilient
to unanticipated shocks, in particular stemming from changes in economic circumstances. Another recent
study (Heutel (2012)) also finds that optimal carbon taxes and emissions are procyclical with business cycles,
implying that carbon pricing mechanisms should respond accordingly to economic fluctuations and cycles.
This paper seeks to contribute to this debate by analyzing the optimal design of and contrasting the
relative performance of two polar instruments - price and quantities - over the business cycles. By doing so,
this paper also links the price-quantity literature with the recent emerging literature that investigates the
optimal design of environmental policies over business cycles. We focus on price-based and quantity-based
policies, most frequently contrasted in the literature (Weitzman (1974) and his many extensions), but we
acknowledge that it is possible to form hybrid instruments, which are a combination of price and quantity
21 Correia
(1996) provide examples in which an incomplete tax system results in non-zero capital-income taxation. See also
discussion in Aruoba and Chugh (2010). de Miguel and Manzano (2006), for instance, show that governments use oil taxes to
accomodate business cycle shocks, if it does not have enough available fiscal instruments (that is under incomplete tax system)
in a small open economy that imports oil.
22 Diamond and Mirrlees, 1971) points out the desirability of undistorted production decisions. The theorem suggests that pure
revenue raising is best done with low rates on large-base taxes, such as VAT or labor taxes. This has important implications for
the potential of the “double dividend” phenomenon associated with environmental policies - benefits additional to the correction
of an environmental market failure - e.g., lower unemployment and/or higher GDP. The second benefit is understood to arise
from the use of energy tax revenues to reduce distortionary taxes elsewehere in the economy. But since increases in energy taxes
lead to tax erosion of the bases of pre-existing labor or capital taxes, in order to raise the same revenue, a higher tax burden is
paid. The benefits from cutting distortionary taxes do not normally overweight the distortion created by the tax erosion effect
and the double dividend arises only in specific circumstances only. For more discussion of the conditions under which a double
dividend arises, see, e.g., Goulder (2013).
20
mechanisms, which are superior to the sole use of either policies considered. We focus on the simplest form
of a cap-and-trade mechanism when considered quantity based regulation, and in particular, abstracted from
so-called banking or borrowing, which in a dynamic setting, can make quantity policies more flexible. We
have analyzed both state-contingent and “basic” fixed priced and fixed based quantity regulation. We find
that the dynamics of the marginal costs are such that they tend to increase during booms and to decline
during recessions making a price instrument preferred to a quantity instrument. We also find that a carbon
tax is essentially constant over the business cycle. Our results thus provide an additional argument and lend
further support to the findings of Pizer (1999), Hoel and Karp (2002) and others who argue in favor of a
price rather than quantity instrument in controlling CO2 emissions in the short-run, when damages from
climate changes remains relatively “flat”.
21
7
Appendix: graphs
output
emissions
1.5
tax
on emissions
−5
pollution stock
1
0.01
3
1
2
0.5
0.005
0.5
0
0
x 10
1
50
100
0
0
fraction of em. abated
1.5
50
100
0
0
abatement spending
4
1
50
100
0
0
gov. bonds
50
100
corporate tax
2
1
1
0
0
−1
2
0.5
0
0
50
100
0
0
labor tax
50
100
−1
0
labor supply
2
50
100
50
100
wage
0.2
1.5
1
1
baseline tax policy
baseline quantity policy
0
0
−1
0
−2
0
0.5
50
100
−0.2
0
50
100
0
0
50
100
Figure 1: Impulse response under baseline carbon tax and cap-and-trade policies to a positive TFP shock
output
emissions
1.5
tax on emissions
pollution stock
1
−5
0.01
3
1
2
0.5
0.005
0.5
0
0
x 10
1
50
100
0
0
fraction of em. abated
50
100
0
0
abatement spending
1
3
0.5
2
50
100
0
0
gov. bonds
50
100
corporate tax
2
1
0
1
0
−0.5
0
1
50
100
0
0
labor tax
−1
50
100
0
0
labor supply
2
50
100
50
1.5
baseline tax policy
fixed tax policy
1
0
0
0.5
50
100
−0.2
0
100
wage
0.2
1
−1
0
−2
0
50
100
0
0
50
100
Figure 2: Responses under baseline carbon tax and fixed carbon tax policies to a positive TFP shock
output
pollution stock
emissions
1.5
fraction of em. abated
−3
1
10
x 10
2
1
1
0.5
5
0.5
0
0
0
50
100
0
0
abatement spending
50
100
0
0
corporate tax
50
100
−1
0
gov. bonds
0
0.1
0.5
2
−0.2
0
0
50
100
−0.4
0
labor supply
50
100
50
100
−0.5
0
50
100
wage
0.2
1.5
1
baseline quantity policy
fixed quantity restriction
0
0.5
−0.2
0
−0.1
0
100
labor tax
4
0
0
50
50
100
0
0
50
100
Figure 3: Responses under baseline cap-and-trade and fixed quantity restriction policies to a positive TFP shock
abatement spending
fraction of emissions abated
4
1
2
0.5
0
0
20
60
80
100
0
0
20
tax on emissions
−5
3
40
x 10
40
60
80
100
80
100
abatement shock
1
2
0.5
1
0
0
20
40
60
80
100
0
0
20
40
60
baseline tax policy
baseline tax policy with abatement cost shock (ρab = 0.4)
baseline tax policy with abatement cost shock (ρab = 0.7)
Figure 4: Responses to a TFP shock and to a TFP shock correlated with a shock to abatement technology
References
Archer, David, “Fate of fossil fuel CO2 in geologic time,” Journal of Geophysical Research: Oceans, 2005,
110 (C9), n/a–n/a.
Aruoba, Unknown AuthorS.B. and S.K. Chugh, “Unknown TitleOptimal fiscal and monetary policy
when money is essential,” Unknown JournalJournal of Economic Theory, 2010, 145, 1618–1647.
Barrage, Lint, “Optimal dynamic carbon taxes in a climate-economy model with distortionary fiscal policy,” working paper, University of Maryland, 2014.
Chari, V. V. and Patrick Kehoe, “Optimal Fiscal and Monetary Policy,” Research Department Staff
Report 251, Federal Reserve Bank of Minneapolis, 1998.
Congressional Budget Office, Effects of a Carbon Tax on the Economy and the Environment, Congress
of the United States, 2013.
Correia, I., “Should Capital Income be Taxed in the Steady State?,” Journal of Public Economics, 1996,
60, 147–151.
Council of Economic Advisers, Economic Report of the President, United States Government Printing
Office, 2013.
Creyts, Jon, Anton Derkach, Scott Nyquist, Ken Ostrowski, and Jack Stephenson, Reducing
U.S. Greenhouse Gas Emissions: How Much at What Costs, U.S. Greenhouse Gas Abatement Mapping
Initiative, 2007.
de Miguel, Carlos and Baltasar Manzano, “Optimal oil taxation in a small open economy,” Review of
Economic Dynamics, 2006, 9, 438–454.
Dessler, Andrew and Edward A. Parson, The Science and Politics of Global Climate Change: A Guide
to the Debate, Cambridge University Press, 2010.
Diamond, Peter A. and James Mirrlees, “Optimal taxation and public production I: production efficiency,” American Economic Review, 1971, 61 (1), 8–27.
Doda, Baran, “How to price carbon in good times... and bad,” Grantham Research Institute on Climate
Change and the Environment, Policy Brief, December, 2014.
Fell, Harrison, Ian A. MacKenzie, and William A. Pizer, “Prices versus quantities versus bankable
quantities,” Resource and Energy Economics, 2012, 34, 607–623.
Gale, William G., Samuel Brown, and Fernando Saltiel, Carbon Taxes as Part of the Fiscal Solution
2013.
Golosov, Mikhail, John Hassler, Per Krusell, and Aleh Tsyvinski, “Optimal Taxes on Fossil Fuel
in General Equilibrium,” Econometrica, 2014, 82 (1), 41–88.
, Maxim Troshkin, Aleh Tsyvinski, and Matthew Weinzierl, “Preference heterogeneity and optimal income taxation,” Journal of Public Economics, 2013, 97, 160–175.
Goulder, Lawrence H., “Climate change policy’s interactions with the tax system,” Energy Economics,
2013, 40, S3–S11.
and Andrew Schein, “Carbon taxes vs. cap and trade: a critical review,” NBER N 19338, 2013.
Hepburn, Cameron, “Regulation by prices, quantities, or both: a review of instrument choice,” Oxford
Review of Economic Policy, 2006, 22 (2), 226–247.
Heutel, Garth, “How Should Environmental Policy Respond to Business Cycles? Optimal Policy under
Persistent Productivity Shocks,” Review of Economic Dynamics, April 2012, 15 (2), 244–264.
Hoel, M. and L. Karp, “Taxes versus quotas for a stock pollutant,” Resource and Energy Economics,
2002, 24, 367–384.
Intergovernmental Panel on Climate Change, Climate Change 2001: The Scientific Basis, Cambridge
University Press, 2001.
, Climate Change 2007: Synthesis Report 2007.
, Working group I contribution to the IPCC Fifth Assessment Report: summary for policymakers 2013.
Karp, L. and J. Zhang, “Regulation of stock externalities with correlated abatement costs,” Environmental
and Resource Economics, 2005, 32, 273–299.
Laffont, Jean Jacques, “More on Prices vs. Quantities,” Review of Economic Studies, 1977, 44(1), 177–
182.
Lintunen, Jussi and Lauri Vilmi, “On optimal emission control Taxes, substitution and business cycles,”
Research Discussion Papers 24/2013, Bank of Finland Oct 2013.
Newell, Richard G. and William A. Pizer, “Regulating stock externalities under uncertainty,” Journal
of Environmental Economics and Management, 2003, 45, 416–432.
Nordhaus, William D., “To slow or not to slow: the economics of the greenhouse effect,” The Economic
Journal, 1991, 101(407), 920–937.
, A question of balance, Yale University Press, 2008.
Parsons, John E. and Luca Taschini, “The role of stocks and shocks concepts in the debate over price
versus quantity,” Environmental and Resource Economics, 2013, 55, 71–86.
Pigou, Arthur Cecil, The economics of welfare, London: Macmillan, 1920.
Pizer, William A., “The optimal choice of climate change policy in the presence of uncertainty,” Resource
and Energy Economics, August 1999, 21 (3-4), 255–287.
, “Combining price and quantity controls to mitigate global climate change,” Journal of Public Economics,
2002, 85(3), 409–34.
, “Climate change catastrophes,” RFF, discussion paper 03-31, 2003.
Rezai, Armon and Frederick van der Ploeg, “Intergenerational inequality aversion, growth and teh
role of damages: Occam’s rule for the global carbon tax,” CEPR discussion paper No. 10292, 2014.
Roberts, M. J. and M. Spence, “Effluent charges and licences under uncertainty,” Journal of Public
Economics, 1976, 5, 193–208.
Rogerson, Richard and Johanna Wallenius, “Micro and macro elasticities in a life cycle model with
taxes,” Journal of Economic Theory, 2009, 144, 2277–2292.
Schmitt-Grohe, Stephanie and Martin Uribe, “Optimal simple and implementable monetary and fiscal
rules,” Journal of Monetary Economics, September 2007, 54 (6), 1702–1725.
Stavins, Robert N., “Correlated uncertainty and policy instrument choice,” Journal of environmental
economics and management, 1996, 30, 218–232.
United States Environmental Protection Agency, Draft inventory of U.S. greenhouse gas emissions
and sinks:1990-2012, National Service Center for Environmental Publications, 2013.
U.S.Census Bureau, Pollution Abatement Costs and Expenditures: 2005, 2008.
Weitzman, Martin L., “Prices vs quantities,” Review of Economic Studies, 1974, 45(2), 229–238.
, “On modeling and interpreting the economics of catastrophic climate change,” The Review of Economics
and Statistics, 2009, 91 (1), 1–19.
, “Can negotiating a uniform carbon price help to internalize teh Global Warming externality?,” working
paper, 2014.
8
Online Appendix: not for publication
8.1
Productivity shock and associated increase in the stock of pollution
In our baseline model a 1% TFP shock results in an increase in the pollution stock of about 0.008% over 25
years. How does this number relate to reality? The Mauna Loa Observatory23 provides monthly information
on the concentration of the atmospheric carbon dioxide. The concentrations are expressed in parts per million
(ppm), which give the ratio of the number of greenhouse gas molecules to the total number of molecules
of dry air. The Carbon Dioxide Information Analysis Center24 provides conversion tables that enable us
to convert this measure of atmospheric CO2 concentration into gigatons of carbon. 1 ppm by volume of
atmosphere CO2 equals to 2.13 GtC. This measure does not count the mass of oxygen in the CO2 molecule,
but since the atom weight of carbon (12 units) and of CO2 (44 units), one unit of GtC is equivalent to
44/12=3.67 GtCO2 (see Dessler and Parson, 2010, p.201), and 1ppm is therefore equivalent to 2.13 GtC and
7.82 GtCO2.
As of July 2014, the concentration of CO2 in atmosphere stood at 399.00 ppm or equivalently at 849.87
GtC or 3059.53 GtCO2. If we treat this value as our steady state, an additional increase of 0.008% in
the pollution stock over 25 years time period, as suggested by impulse response function, corresponds to
0.032 ppm, or 0.068 GtC and 0.24 GtCO2 increase in the concentration of CO2 in atmosphere. World CO2
emissions in 2012 stood at 34.5 GtCO2. Assuming that the level of yearly emissions does not change, over
a period of 25 years the world will emit 862.5 GtCO2 meaning that an additional increase in the CO2 stock
due to a TFP shock constitutes only 0.028% of all emissions over a 25-year period.
Intergovernmental Panel on Climate Change (2007) reports a table (Table 5.1, p. 67) that relates
CO2 concentration in the atmosphere to the global temperature and average sea level increase above preindustrial levels. At CO2 concentrations of 350-400 ppm (current level), global temperature increase above
pre-industrial levels ranges from 2.0-2.5◦ C, and the global average sea level rises above the pre-industrial
level from 0.4-1.4 m. For the 400-440 ppm range the corresponding numbers are: 2.4-2.8◦ C and 0.5-1.7◦ C.
Thus an increase in CO2 concentration from the current level to 440 ppm (by 40 ppm) could lead to a
maximal increase in the temperature above pre-industrial levels of 0.4◦ C (2.8◦ C-2.4◦ C) and the maximum
sea rise level of 0.3 m (1.7m-1.4m). Treating these estimates as our reference, we can conclude that an
additional increase in CO2 concentrations of 0.032 ppm would correspond to an increase in temperature by
0.0003192◦ C and an additional increase in the sea level by 0.0002394 m over 25 years interval following the
productivity shock.
23 http://co2now.org/
24 http://cdiac.ornl.gov/pns/convert.html
8.2
Appendix A: Data sources
In this section we describe data sources and USA data entry components into table 2.
Data from the NIPA tables are for year 2013.
• GDP - from the NIPA Table 1.5.5. Gross Domestic Product, Expanded Detail, line 1.
• Personal consumption expenditure - from the NIPA Table 1.5.5. Gross Domestic Product, Expanded
Detail, line 2.
• Government consumption expenditure - from the NIPA Table 1.5.5. Gross Domestic Product, Expanded Detail, line 55+line 58+line 61.
• Government gross investment - from the NIPA Table 1.5.5. Gross Domestic Product, Expanded Detail,
line 56+line 59+line 62.
• Gross private domestic investment - from the NIPA Table 1.5.5. Gross Domestic Product, Expanded
Detail, line 26.
• Emissions per unit of total GDP - for 2012 for the US from the United States Environmental Protection
Agency (2013), p.ES-24, Table ES-9.
• Fraction of emissions abated: derived from author’s calculations with original data from Creyts et
al. (2007), who provide estimates of potential abatement projections for greenhouse gases in the US.
They estimate that the US would potentially abate cumulative 3GtCO2 of emissions for the period
2005-2030. Assuming the same amount of emissions abated every year during 25 years time period,
from 2005 to 2030, and given that total greenhouse gas emissions amounted to 6.5GtCO2 by the US
in 2012 (United States Environmental Protection Agency (2013)), we obtain 1.85%, an estimate of the
fraction of emissions abated in 2012.
• Abatement Spending - from the U.S.Census Bureau (2008), Table 1 (Pollution Abatement Operating
Costs) and Table 2 (Pollution Abatement Capital Expenditures). U.S.Census Bureau (2008) is a survey
of a sample of 20000 manufacturing plants, which, according this survey, spent 20677.6 mln USD on
pollution abatement operating costs and 5907.8 mln USD on pollution abatement capital expenditures
in 2005. By combining these data with the US GDP data for 2005, USD 13095.4 bln, we obtain
estimate of the fraction of abatement spending in GDP, 0.2%, reported in the main part of the paper.
• Labor tax - OECD, Taxing Wages 2014 (May 2014), http://www.oecd-ilibrary.org/taxation/
taxing-wages-2014_tax_wages-2014-en
• Central government corporate income tax rate - OECD, Taxation of Corporate and Capital Income,
Corporate Income Tax
• Revenue from environmental taxes - Congressional Budget Office (2013) estimates potential tax revenues from carbon taxes at 1.2 trillion USD in a 10 years period. Assuming a yearly revenue of 0.12
trillion USD, we calculate it as a fraction of US GDP in 2013 and obtain the estimate 0.7% of GDP.
• Steady state value of government bonds as relation to output - based on Table B79 (federal debt held
by public as percent of gross domestic product) from Council of Economic Advisers (2013).
8.3
Appendix B: First-order conditions of the Ramsey problem
The first-order conditions of the Ramsey problem outlined in section 2.6 are given by:
u0c (t) − λt u00cc (t) + λt−1 u00cc (t)(1 − δ + rt ) + Ωt − (Λt τLt lt − ςt (1 + τLt ))
λt−1 u0c (t)
u00Lc (t)u0c (t) − u0L (t)u00cc (t)
= 0 (35)
(u0c (t))2
∂rt
+ Ωt m0 (µ)yt − χt yt m00 (µ) − Λt τEt h(yt ) + ςt (1 − d(xt ))fL0 [τEt h0 (yt ) − m0 (µt )] + Φt h(yt ) = 0 (36)
∂µt
∂rt+1
+ Ωt − β(1 − δ)Ωt+1 − λpt+1 β(1 − d(xt+1 ))fk0 (t + 1) +
∂kt
+βςt+1 (1 − d(xt+1 ))fkL (t + 1)[1 − τct+1 − τEt+1 (1 − µt+1 h0 (yt+1 ) − m(µt+1 ] = 0
(37)
∂rt
+ Ωt (m(µt ) − 1) + χt [τEt h0 (yt ) − m0 (µt )] + Λt [τEt (1 − µt )h0 (yt ) + τct ] +
∂yt
+λpt + ςt (1 − d(xt ))fL0 [−τEt (1 − µt )h00 (yt )] − Φt (1 − µt )h0 (yt ) = 0
(38)
λt βu0c (t + 1)
λt−1 u0c
λt−1 u0c
∂rt
+ λpt d0 (xt )f (t) − ςt d0 (xt )fL0 [1 − τct − τEt (1 − µt )h0 (yt ) − m(µt )] + Φt − βηΦt+1 = 0
∂xt
u0L − λt u00cL + λt−1 u00cL (1 − δ + rt ) + λt−1 u0c
(39)
∂rt
+
∂lt
u00 u0 − u0 u00
u00 u0 − u0 u00
u0L
τLt − τLt lt Lc c 0 2 L cc ] − λpt (1 − d(xt ))fL0 + ςt (1 + τLt ) Lc c 0 2 L cc +
0
uc
(uc )
(uc )
00
0
+ςt (1 − d(xt ))fLL [1 − τct − τEt (1 − µt )h (yt ) − m(µt )] = 0
+Λt [−
−Λt
λt−1 u0c
u0L
u0L
l
+
ς
=0
t
t
u0c
u0c
(40)
(41)
∂rt
+ χt h(yt ) + Λt (1 − µt )h(yt ) − ςt (1 − d(xt ))fL0 (1 − µt )h0 (yt ) = 0
∂τEt
(42)
∂rt
+ Λt yt − ςt (1 − d(xt ))fL0 = 0
∂τct
(43)
λt−1 u0c
Λt ρBt − βΛt+1 = 0
u0g − λt u00cg + λt−1 u00cg (1 − δ + rt ) − Λt + Ωt + (ςt (1 + τLt ) − Λt τLt lt )
(44)
u00Lg u0c − u0L u00cg
=0
(u0c )2
(45)
output
emissions
1.5
tax on emissions
pollution stock
1
0.01
0.5
0.005
−5
3
1
2
0.5
0
0
1
50
100
0
0
fraction of em. abated
50
100
0
0
abatement spending
50
100
0
0
gov. bonds
3
2
1
0.5
2
1
0
0
1
0
−1
50
100
0
0
labor tax
50
100
−1
0
labor supply
4
50
100
0.5
−0.5
0
50
100
shock persistence ρ=0.95
shock persistence ρ=0.6
shock persistence ρ=0.4
0
0
100
−2
0
1.5
1
50
100
wage
0.5
2
−2
0
50
corporate tax
1
−0.5
0
x 10
50
100
0
0
50
100
Figure 5: Responses to a TFP shock under different values of the persistence of the shock ρ